Hi @alfBI ,
It does seem unusual for Power BI models and Gen1 dataflows to consume such a significant portion of your daily capacity—especially if they are only referencing existing corporate Gen1 dataflows and Excel files hosted on SharePoint. While some CU (Capacity Unit) consumption is expected during refresh and transformation processes, usage at the level you describe (15% for models and 5% for DF1) suggests that there may be inefficiencies in how the data is being processed. It's quite possible that the queries or transformations are overly complex, poorly optimized, or unnecessarily duplicating work during refreshes.
Moving your corporate Gen1 dataflows to a Fabric Lakehouse could indeed lead to performance improvements, as Lakehouse provides a more scalable and performance-optimized architecture for large datasets. However, it would still be wise to conduct a thorough review of the models and dataflows in question. Look for issues like repeated joins, excessive calculated columns, use of nested functions in Power Query, or inefficient data types. Optimizing these aspects could reduce CU consumption significantly, regardless of the backend. Transitioning to a Lakehouse is a strategic move, but ensuring best practices in model and dataflow design is equally critical to sustained performance gains.
Did it work? ✔ Give a Kudo • Mark as Solution – help others too!
